S Series Switch. QoS Technology White Paper. Issue 01. Date HUAWEI TECHNOLOGIES CO., LTD.

Transcription

1 Issue 01 Date HUAWEI TECHNOLOGIES CO., LTD.

2 2013. All rights reserved. No part of this document may be reproduced or transmitted in any form or by any means without prior written consent of Huawei Technologies Co., Ltd. Trademarks and Permissions and other Huawei trademarks are trademarks of Huawei Technologies Co., Ltd. All other trademarks and trade names mentioned in this document are the property of their respective holders. Notice The purchased products, services and features are stipulated by the contract made between Huawei and the customer. All or part of the products, services and features described in this document may not be within the purchase scope or the usage scope. Unless otherwise specified in the contract, all statements, information, and recommendations in this document are provided "AS IS" without warranties, guarantees or representations of any kind, either express or implied. The information in this document is subject to change without notice. Every effort has been made in the preparation of this document to ensure accuracy of the contents, but all statements, information, and recommendations in this document do not constitute a warranty of any kind, express or implied. Huawei Technologies Co., Ltd. Address: Website: Huawei Industrial Base Bantian, Longgang Shenzhen People's Republic of China i

3 Contents Contents 1 Introduction to QoS What Is QoS? QoS Specifications Bandwidth/Throughput Delay Delay Variation (Jitter) Packet Loss Rate Common QoS Specifications QoS Service Models Best-Effort IntServ DiffServ Comparison Between DiffServ and IntServ Models Components in the DiffServ Model Traffic Classification and Marking Simple Traffic Classification Complex Traffic Classification Traffic Marking Application of Traffic Classification and Marking Traffic Policing and Traffic Shaping Traffic Policing What Is a Token Bucket CAR Traffic Shaping Comparison Between Traffic Policing and Traffic Shaping Congestion Management and Congestion Avoidance Background Congestion Management Congestion Avoidance Application Scenarios User-based Differentiated Services ii

4 Contents Networking Requirements Configuration Roadmap Procedure Service-based Differentiated Services Networking Requirements Configuration Roadmap Procedure Troubleshooting Cases Packets Enter Incorrect Queues Priority Mapping Results Are Incorrect Traffic Policy Does Not Take Effect FAQ Does the S9700 Collect Traffic Statistics Based on Packets or Bytes? What Are the Differences Between Interface-based CAR and Global CAR? How Does Level-2 CAR Take Effect? A Traffic Policy Contains an ACL Rule Defining TCP or UDP Port Number Range. When the Traffic Policy Is Delivered, the System Displays the Message "Add rule to chip failed." Why? CAR Is Incorrect. Why? An ACL Applied to the Outbound Direction Cannot Define the Port Number Range. Why? Can 802.1p Re-marking and Traffic Statistics Be Configured in a Traffic Policy Simultaneously on the S9700? When Both QinQ and Traffic Policy-based VLAN Stacking Are Configured on an Interface, Which Configuration Takes Effect? Why ACL Rule Update May Cause Instant Traffic Interruption? After an ACL or QoS Is Configured, the Configuration Is Invalid for Mirroring Packets. Why? Why a Traffic Policy Containing Traffic Filtering or CAR Is Invalid for Incoming Packets on an S9700? Why PQ+DRR Configured on an S9700 Interface Does Not Take Effect? Why Priorities in Outgoing Mirroring Packets Are Not Changed After Priority Mapping Is Configured? When You Configure a Deny Rule in a Traffic Policy Containing Flow Mirroring, Normal Service Traffic Is Affected. Why? When a Traffic Policy Containing Flow Mirroring Is Applied to an Interface, the Global Traffic Policy Becomes Invalid. Why? What Is the Relationship Between an ACL and a Traffic Policy? How Are Packets Forwarded Using PBR on S Series Switches? Appendix Common Service Priorities Port Numbers of Common Application Services Common Queue Scheduling Solution Recommended WRED Parameter Setting Color-based WRED Parameter Setting Queue-based WRED Parameter Setting Video Service Bandwidth Usage Coding-based Video Bandwidth iii

5 Contents HD-based Video Bandwidth Video Conference Bandwidth Audio Bandwidth Usage Audio Bandwidth Based on Codec Technologies iv

6 1 Introduction to QoS 1 Introduction to QoS 1.1 What Is QoS? As network technologies rapidly develop, services on the Internet become increasingly diversified. Apart from traditional applications such as WWW, , and File Transfer Protocol (FTP), the Internet has expanded to encompass other services such as IP phones, e-commerce, multimedia games, e-learning, telemedicine, videophones, videoconferencing, video on demand (VoD), and online movies. In addition to web page browsing, new enterprises require services including identity authentication of employees and visitors, remote video conferencing, s, video, FTP file upload and download, and Telnet services on special devices in working hours. These new services have special requirements on the bandwidth, delay, and delay variation. For example, videoconferencing and VoD services demand high bandwidth, short delay, and low delay variation. Key tasks such as transaction processing and Telnet require short delay and preferential handling when congestion occurs, although such tasks do not necessarily demand high bandwidth. Figure 1-1 Enterprise new services Data server Videoconferencing Videoconferencing Internet VoD VoD Enterprise branch 1 Enterprise branch 2 Diversified services enrich people's lives but also increase the risk of traffic congestion on the Internet. When traffic congestion occurs, services encounter long delays or even packet loss. As a result, services deteriorate or even become unavailable. Therefore, a solution to resolve traffic congestion on the IP network is urgently needed. 1

7 1 Introduction to QoS The best way to limit traffic congestion is to increase network bandwidths. However, increasing network bandwidths is not feasible due to the high operation and maintenance costs. The most cost-effective way is to use a "guarantee" policy to management traffic congestion. This method is quality of service (QoS). QoS provides end-to-end service guarantee for differentiated services and has played an overwhelmingly important role on the Internet. Without QoS, service quality cannot be guaranteed. 1.2 QoS Specifications QoS provides customized service guarantee for key services based on the following specifications: Bandwidth/Throughput Delay Delay variation (jitter) Packet loss rate Bandwidth/Throughput Bandwidth, also called throughput, refers to the maximum number of bits allowed to transmit between two ends within 1 second or the average rate at which specific data flows are transmitted between two network nodes. Bandwidth is expressed in bit/s. The water supply network is used to help you understand bandwidth. The diameter of a water supply pipe measures the capability to carry water. The diameter of the water supply pipe is similar to bandwidth and water is similar to data. A thick pipe indicates higher bandwidth and greater capability to transmit data. As services become increasingly diversified, Internet citizens expect higher bandwidths so they can not only browse the Internet for news but also experience any number of popular applications. The epoch-making information evolution continually delivers new and attractive applications, such as new-generation multimedia, video transmission, database, and IPTV, all of which demand extremely high bandwidths. Therefore, bandwidth is always the major focus of network planning and provides an important basis for network analysis. Figure 1-2 Insufficient bandwidth Have you watched video I have recommended? The Internet is slow, and the online video cannot play smoothly. Maybe I have to download it. IP network 2

8 1 Introduction to QoS Two concepts, upstream rate and downstream rate, are relevant to bandwidth. The upstream rate refers to the rate at which users send information to the network, and the downstream rate refers to the rate at which the network sends data to users. For example, the rate at which users upload files to the network through FTP is determined by the upstream rate, and the rate at which users download files is determined by the downstream rate Delay A delay refers to the period of time during which a packet is transmitted from a source to its destination. Use voice transmission as an example. A delay refers to the period during which words are spoken and then heard. If the delay is too long, voices become unclear or interrupted. Most users are insensitive to a delay of less than 100 ms. If a delay ranging from 100 ms to 300 ms occurs, the speaker can sense slight pauses in the responder's reply, which can seem annoying to both. If a delay greater than 300 ms occurs, both the speaker and responder sense an obvious delay and have to wait for responses. If the speaker cannot wait but repeats what has been said, voices overlap, and the quality of the conversation deteriorates severely. Figure 1-3 Long delay Hello! (2s later) Can you hear me? IP network (2s later) Hello! (4s later) Is that Jack? Interrupted for so long? Delay Variation (Jitter) Jitter refers to the difference in delays of packets in the same flow. If the period before a packet that has reached a device is sent by the device differs from one packet to another in a flow, jitters occur, and service quality is affected. Specific services, especially voice and video services, are zero-tolerant of jitters, because jitter will interrupt voice or video services. 3

9 1 Introduction to QoS Figure 1-4 High jitter I know, you don't. I know you don't? IP network Time I know, you don't. D3=10 ms D4=40 ms D5=90 ms D2=50 ms D1=50 ms I know you D6 = 90 ms don't Packet Loss Rate Jitters also affect protocol packet transmissions. Specific protocol packets are transmitted at a fixed interval. If high jitters occur, such protocols flap, adversely affecting quality. Jitter thrives on networks but service quality will not be affected if jitters do not exceed a specific tolerance. Buffers can alleviate excess jitters but prolong delays. Packet loss occurs when one or more packets traveling across a network fail to reach their destination. Slight packet loss does not affect services. For example, users are unaware of the loss of a bit or a packet in voice transmission. If a bit or a packet is lost in video transmission, the image on the screen becomes momentarily garbled but the image recovers very quickly. Even if TCP is used to transmit data, slight packet loss is not a problem because TCP instantly retransmits the packets that have been lost. If severe packet loss does occur, packet transmission efficiency is affected. The packet loss rate indicates the severity of service interruptions on networks and concerns users. Figure 1-5 High packet loss rate I have sent a file to you. I have sent you What? IP network 4

10 1 Introduction to QoS 1.3 Common QoS Specifications Internet users have different requirements for the bandwidth, delay, jitter, and packet loss rate for different services on the IP network. Table 1-1 and Table 1-2 list QoS specifications for different services. Table 1-3 list QoS specifications defined by the Metro Ethernet Forum (MEF), including availability, delay, jitter, loss, and restoration time. Table 1-1 QoS specifications for common services Enterprise Service Type Bandwidth/Throughput Delay Jitter Packet Loss Rate Videoconferencing High Very low Very low Low and predictable E-commerce Medium Low Low Low Streaming media High Low Low Low and predictable s and file transfer Low Not important Not important Not important HTML web page browsing Not specific Medium Medium Not important FTP client/server Medium Low Low Low Table 1-2 Reference values of QoS specifications for common services Enterprise Service Type Delay Jitter Packet Loss Rate Videoconferencing 50 ms 10 ms 0.1% E-commerce 200 ms 100 ms TCP guarantee Streaming media 1s 200 ms 0.1% s and file transfer N/A N/A TCP guarantee HTML web page browsing N/A N/A NA FTP client/server N/A N/A TCP guarantee 5

11 1 Introduction to QoS Table 1-3 QoS specifications defined by the MEF Service Class Service Characteristics Service Performance Premium Silver Bronze Real-time IP telephony or IP video applications Burst mission-critical data applications requiring low loss and delay such as storage Burst data applications requiring bandwidth assurances Availability > 99.99% Delay < 40 ms Jitter < 1 ms Loss < 0.1% Restoration time: 50 ms Availability > 99.99% Delay < 50 ms Jitter: N/A Loss < 0.1% Restoration time: 200 ms Availability > 99.90% Delay < 500 ms Jitter: N/A Loss: N/A Restoration time: 2s Standard Best effort service Availability > 97.00% Delay: N/A Jitter: N/A Loss: N/A Restoration time: 5s 6

12 2.1 QoS Service Models Best-Effort IntServ Network applications require end-to-end communication. Traffic may traverse multiple switches on one network or even multiple networks before reaching the destination host. Therefore, to provide end-to-end QoS guarantee, an overall network deployment is required. Service models are used to provide an end-to-end QoS guarantee based on specific requirements. QoS provides the following types of service models: Best-Effort Integrated service (IntServ) Differentiated service (DiffServ) Best-Effort is the default service model on the Internet and applies to various network applications, such as FTP and . It is the simplest service model. Without network notification, an application can send any number of packets at any time. The network then makes its best attempt to send the packets but does not provide any guarantee for performance such as delay and reliability. The Best-Effort model applies to services that that do not require low delay and high reliability. Before sending a packet, IntServ uses signaling to apply for a specific level of service from the network. The application first notifies the network of its traffic parameters and specific service qualities, such as bandwidth and delay. After receiving a confirmation that sufficient resources have been reserved, the application sends the packets. The network maintains a state for each packet flow and executes QoS behaviors based on this state to fulfill the promise made to the application. The packets must be controlled within the range described by the traffic parameters. 7

13 IntServ uses the Resource Reservation Protocol (RSVP) as signaling, which is similar to Multiprotocol Label Switching Traffic Engineering (MPLS TE). RSVP reserves resources such as bandwidth and priority on a known path and each network element along the path must reserve required resources for data flows requiring QoS guarantee. That is, each network element maintains a soft state for each data flow. A soft state is a temporary state and is periodically updated using RSVP messages. Each network element checks whether sufficient resources can be reserved based on these RSVP messages. The path is available only when all involved network elements can provide sufficient resources. Figure 2-1 IntServ model PC Phone OK I require 2 Mbit/s bandwidth. I require 2 Mbit/s bandwidth. OK I require 2 Mbit/s bandwidth. OK OK I require 2 Mbit/s bandwidth. VoIP STB IPTV DiffServ The IntServ model provides end-to-end guarantee, but has the following limitations: MPLS TE is feasible because it is deployed on the core network and the network scale is controllable. The IntServ model involves end-to-end services at the core, aggregation, and access layers, and more network elements. The complex network limits its development. IntServ must be supported by all network nodes. Core, aggregation, and access devices have different performances, and some of them may not support the IntServ model. The IntServ model cannot be widely applied to the Internet backbone network. DiffServ classifies packets on the network into multiple classes for differentiated processing. When traffic congestion occurs, classes with a higher priority are given preference. This function allows packets to be differentiated and to have different packet loss rates, delays, and jitters. Packets of the same class are aggregated and sent as a whole to ensure the same delay, jitter, and packet loss rate. In the DiffServ model, edge nodes classify and aggregate traffic. Edge nodes classify packets based on a combination of fields, such as the source and destination addresses of packets, precedence in the ToS field, and protocol type. Edge nodes also re-mark packets with different priorities, which can be identified by other nodes for resource allocation and traffic control. Therefore, DiffServ is a flow-based QoS model. 8

14 Figure 2-2 DiffServ model Flow Type Priority Voice 0 Video 0 Data 0 Voice flow Data flow Video flow Flow Type Priority Voice 0 Voice flow (priority 5) Video 0 Data 0 Video flow (priority 4) Data flow (priority 0) Video flow Voice flow VoIP IPTV Data flow Data server Different from IntServ, DiffServ requires no signaling. In the DiffServ model, an application does not need to apply for network resources before transmitting packets. Instead, the application notifies the network nodes of its QoS requirements by setting QoS parameters in packets. The network does not need to maintain a state for each data flow but provides differentiated services based on the QoS parameters of each data flow. DiffServ classifies incoming packets on the network edge and manages packets of the same class as a whole to ensure the same transmission rate, delay, and jitter. DiffServ processes flows of each type separately. Network edge nodes mark packets with a specific service class in packet headers, and then apply traffic management policies to the packets based on the service class. Interior nodes perform specific behaviors for packets based on packet information. DiffServ takes full advantage of network flexibility and extensibility of the IP network and transforms information in packets into per-hop behaviors, greatly reducing signaling operations. Therefore, DiffServ not only adapts to Internet service provider (ISP) networks but also accelerates IP QoS applications on live networks. It is the mainstream model on networks. 9

15 Entities in the DiffServ Model Figure 2-3 Entities in the DiffServ model PHB-based forwarding Interior node DS domain Boundary node DS domain Interior node Boundary node SLA/TCA Boundary node Boundary node User network Classify and aggregate services Different PHBs in different DSs, being coordinated based on the SLA/TCA User network DS node: a network node that implements the DiffServ function. All network elements in Figure 2-3 are DS nodes. DS domain: a set of contiguous DS nodes that adopt the same service policy and per-hop behavior (PHB). One DS domain covers one or more networks under the same administration. For example, a DS domain can be an ISP's network or an organization's intranet. For an introduction to PHB, see the next section. A PHB describes the externally observable forwarding treatment applied to a DS node. DS boundary node: connects to another DS domain or a non-ds-aware domain. The DS boundary node classifies and manages incoming traffic. DS interior node: connects to DS boundary nodes and other interior nodes in one DS domain. DS interior nodes implement simple traffic classification based on DSCP values, and manage traffic. SLA/TCA: The SLA refers to the services that the ISP promises to provide for individual users, enterprise users, or adjacent ISPs that need intercommunication. The SLA covers multiple dimensions, including the accounting protocol. The service level specification (SLS) provides technique description for the SLA. The SLS focuses on the traffic control specification (TCS) and provides detailed performance parameters, such as the committed information rate (CIR), peak information rate (PIR), committed burst size (CBS), and peak burst size (PBS). DS region: consists of one or more adjacent DS domains. Different DS domains in one DS region may use different PHBs to provide differentiated services. The SLA and traffic conditioning agreement (TCA) are used to allow for differences between PHBs in different DS domains. The SLA or TCA specifies how to maintain consistent processing of the data flow from one DS domain to another. 10

16 PHB An action taken for packets on each DS node is called PHB. PHB is a description of the externally observable forwarding treatment applied to a DS node. You can define PHB based on priorities or QoS specifications such as the delay, jitter, and packet loss ratio. The PHB defines some forwarding behaviors but does not specify the implementation mode. Currently, the IETF defines four types of PHBs: Class Selector (CS), Expedited Forwarding (EF), Assured Forwarding (AF), and best-effort (BE). BE is the default PHB. RFC 2597 classifies AF into four classes: AF1 to AF4. RFC 2474 classifies CS into CS6 and CS7. There are eight types of PHBs. Each PHB corresponds to a Class of Service (CoS) values. Different CoS values determine different congestion management policies. In addition, each PHB is assigned three drop priorities, also called colors (green, yellow, and red). Different drop priorities determine congestion avoidance policies of different flows. For details about CoS values and colors, see Priority Mapping. For details about congestion management and congestion avoidance, see section 2.4 "Congestion Management and Congestion Avoidance." Table 2-1 describes standard PHBs and their usage. Table 2-1 Standard PHBs and usage PHB Description Sub-PHB Usage CS (RFC 2474) The CS PHB indicates the same service class as the IP precedence value. The CS PHB is of the highest priority among standard PHBs. CS7 CS6 CS6 and CS7 PHBs are used for protocol packets by default, such as STP, LLDP, and LACP packets. If these packets are not forwarded, protocol services are interrupted. EF (RFC 2598) The EF PHB defines that the rate at which packets are sent from any DS node must be higher than or equal to the specified rate. The EF PHB cannot be re-marked in the DS domain but can be re-marked on the edge nodes. The EF PHB applies to real-time services that require a short delay, low jitter, and low packet loss rate, such as video, voice, and video conferencing. - EF PHB is used for voice services. Voice services require a short delay, low jitter, and low packet loss rate, and are second only to protocol packets in terms of importance. The bandwidth dedicated to EF PHB must be restricted so that other services can use the bandwidth. AF (RFC 2597) The AF PHB defines that traffic exceeding AF4 AF4 PHB is used for signaling of voice services. 11

17 PHB Description Sub-PHB Usage the specified bandwidth (as agreed to by users and an ISP) can be forwarded. The traffic that does not exceed the bandwidth specification is forwarded as required, and the traffic that exceeds the bandwidth specification is forwarded at a lower priority. The AF PHB applies to services that require a short delay, low packet loss rate, and high reliability, such as e-commerce and VPN services. AF3 AF2 AF1 Signaling is used for call control, during which a seconds-long delay is tolerable, but no delay is allowed during a conversation. Therefore, the processing priority of voice services is higher than that of signaling. AF3 PHB is used for Telnet and FTP services. The services require medium bandwidth and reliable transmission, but are sensitive to the delay and jitter. AF2 PHB is used for live programs of IPTV and ensures smooth transmission of online video services. Live programs are real-time services, requiring continuous bandwidth and a large throughput guarantee. They allow less packet loss. AF1 PHB is used for common data services such as s. Common data services require only zero packet loss, and do not require high real-time performance and jitter. BE (RFC 2474) The BE PHB focuses only on whether packets can reach the destination, regardless of the transmission performance. Any switch must support BE PHB. - BE PHB applies to best-effort services on the Internet, such as HTTP web page browsing services Comparison Between DiffServ and IntServ Models Table 2-2 Comparison between DiffServ and IntServ models Item DiffServ IntServ End-to-end QoS guarantee Network scale Implements end-to-end QoS guarantee by connecting multiple DS domains. Applies to various networks, and applies to large-scale networks using multiple DS domains. Directly implements end-to-end QoS guarantee. Is inapplicable to large-scale networks. 12

18 Item DiffServ IntServ Network cost Network element cost Has no extra cost because DiffServ model notifies other devices of packet priorities using packet precedence fields. Has low cost because resources do not need to be reserved for network elements. Has extra cost because the IntServ model uses RSVP to notify other devices and periodically update network resources. Has high cost because resources need to be reserved for network elements Components in the DiffServ Model The DiffServ model consists of four QoS components: Traffic classification and marking: Traffic classification classifies packets while keeping the packets unchanged. Traffic marking sets different priorities for packets and therefore changes the packets. Traffic policing and shaping: Limit the traffic rate. When traffic exceeds the specified rate, traffic policing drops excess traffic, and traffic shaping buffers excess traffic. Congestion management and avoidance: Congestion management buffers packets in queues when traffic congestion occurs and determines the forwarding order based on a specific scheduling algorithm. Congestion avoidance monitors network resources. When network congestion aggravates, the device drops packets to regulate traffic so that the network is not overloaded. Port mirroring and traffic mirroring: Mirroring copies packets on a specified interface to the mirroring destination interface that is connected to a data monitoring device. Then you can use the data monitoring device to analyze the packets copied to the destination interface, and monitor the network and troubleshoot faults. Traffic classification and marking are the basis for implementing differentiated services. Traffic policing, traffic shaping, congestion management, and congestion avoidance control network traffic and allocated resources. Packets are processed by the components in sequence, as shown in Figure

19 Outbound interface Scheduling Traffic classification Inbound interface Data Voice Video S Series Switch Figure 2-4 Processing of QoS components Enter queue Congestion avoidance Queue 0 Congestion management Traffic shaping Leave queue Traffic policing (CAR) Other processing Queue 1 Queue 2 Queue N The four QoS components are implemented at different locations on a network according to the DiffServ model and service development. Traffic classification, traffic marking, and traffic policing are performed in the inbound direction on an access interface, traffic shaping is performed in the outbound direction on an access interface, and congestion management and congestion avoidance are performed in the outbound direction on a network-side interface. If services with different CoS values are transmitted on an access interface, queue scheduling and a packet drop policy must be configured in the outbound direction on the access interface. 2.2 Traffic Classification and Marking Traffic classification technology allows a device to classify packets that enter a DiffServ domain so that other applications or devices learn about the packet service type and apply any appropriate action upon the packets. Packets can be classified based on QoS priorities, or packet information such as the source IP address, destination IP address, MAC address, IP protocol, and port number, or specifications in an SLA. After packets are classified on the DiffServ domain edge, internal nodes provide differentiated services for the packets that are classified. A downstream node can resume the classification result calculated on an upstream node or perform another traffic classification based on its own criteria. Traffic classification is classified into simple traffic classification and complex traffic classification. For details, see section "Simple Traffic Classification" and section "Complex Traffic Classification." Simple Traffic Classification Simple traffic classification classifies packets based on simple rules, for example, 802.1p priorities in VLAN packets, ToS values in IP packets, TC values in IPv6 packets, EXP values in MPLS packets, to identify traffic with different priorities or CoS values and implement mapping between external and internal priorities. 14

20 QoS Priority Fields Simple traffic classification trusts priorities in upstream packets on an interface and performs priority mapping. That is, simple traffic classification maps QoS priorities in upstream packets to CoS values and colors, and maps CoS values and colors in downstream packets to QoS priorities. Simple traffic classification is deployed on DS interior nodes. DiffServ provides differentiated services for packets that carry different QoS information in specific fields. The fields are described as follows: 802.1p priority VLAN packets are classified based on the 802.1p priority (PRI) in the packets. The PRI field in a VLAN packet header identifies the QoS requirement. The PRI field is 3 bits long and indicates precedence. The value ranges from 0 to 7 with a larger value reflecting a higher precedence. Figure p priority in a VLAN packet 16bits 3bits 1bits 12bits TPID PRI CFI VLAN ID ToS field in an IP packet In an IPv4 packet header, the three leftmost bits (IP precedence) in the ToS field or the six leftmost bits (DSCP field) in the ToS field are used to identify a QoS priority. The IP precedence classifies packets into a maximum of eight classes, and the DSCP field classifies packets into a maximum of 64 classes. Figure 2-6 ToS field in an IPv4 packet header RFC 2474 DSCP field RFC 1349 Precedence D T R C 8 bit Version HeadLength ToS Total Length RFC 1349 defines bits in the ToS field as follows: Bits 0 to 2 refer to the precedence. The value ranges from 0 to 7 with a larger value reflecting a higher precedence. The ToS field in IP packets is similar in function to the 802.1p priority in VLAN packets. The D bit refers to the delay. The value 0 indicates no specific requirement for the delay and the value 1 indicates that the network is required to minimize the delay. 15

21 The T bit refers to the throughput. The value 0 indicates no specific requirement for the throughput and the value 1 indicates that the network is required to maximize the throughput. The R bit refers to reliability. The value 0 indicates no specific requirement for reliability and the value 1 indicates that the network demands high reliability. The C bit refers to the monetary cost. The value 0 indicates no specific requirement for the monetary cost and the value 1 indicates that the network is required to minimize the monetary cost. Bits 6 and 7 are reserved. RFC 2474 defines bits 0 to 6 as the DSCP field, and the three leftmost bits indicate the class selector code point (CSCP) value, which identifies a class of DSCP. The DSCP value ranges from 0 to 7 with a larger value reflecting a higher precedence. The DSCP value in IP packets is similar in function to the 802.1p priority in VLAN packets. The three rightmost bits are seldom used and are not mentioned here. EXP field in an MPLS packet header Multiprotocol Label Switching (MPLS) packets are classified based on the EXP field value. The EXP field is 3 bits long and indicates precedence. The value ranges from 0 to 7 with a larger value reflecting a higher precedence. The EXP field in MPLS packets is similar in function to the ToS field or DSCP field in IP packets. Figure 2-7 EXP field in an MPLS packet header 20bits 3bits 1bits 8bits Label Exp S TTL Table 2-3 describes the mapping from the IP precedence, EXP, and 802.1p values to the DSCP value. Table 2-3 Mapping from the IP precedence, EXP, and 802.1p values to the DSCP value IP Precedence MPLS EXP Value 802.1p Priority DSCP Value Table 2-4 describes the mapping from the DSCP value to 802.1p, EXP, and IP precedence values. 16

22 Table 2-4 Mapping from the DSCP value to 802.1p, EXP, and IP precedence values DSCP Value IP Precedence MPLS EXP Value 802.1p Priority Priority Mapping The priority field in a packet varies with network type. For example, a packet carries the 802.1p field on a VLAN, the DSCP field on an IP network, and the EXP field on an MPLS network. To provide differentiated services for different packets, the switch maps the QoS priority of incoming packets to the CoS value (also called scheduling precedence) and drop precedence (also called color), and then performs congestion management based on the CoS value and congestion avoidance based on the color. Before forwarding packets out, the device maps the CoS value and color back to the QoS priority so that other devices can process the packets based on the QoS priority. A device maps the QoS priority to the CoS value and color for incoming packets and maps the CoS value and color back to the QoS priority for outgoing packets, as shown in Figure 2-8. Figure 2-8 Priority mapping Upstream 802.1p DSCP MPLS EXP ATM service type & CLP Mapping CoS Color SFU CoS Mapping 802.1p DSCP Color MPLS EXP Downstream ATM CLP 17

23 CoS The CoS refers to the internal service class of packets. Eight CoS values are available: Class Selector 7 (CS7), CS6, Expedited Forwarding (EF), Assured Forwarding 4 (AF4), AF3, AF2, AF1, and Best Effort (BE). CoS determines the type of queues to which packets belong. The priority of queues with different CoS values depends on the scheduling algorithms used: If queues with eight CoS values all use priority queuing (PQ), the priority of queues is: CS7 > CS6 > EF > AF4 > AF3 > AF2 > AF1 > BE. If the BE queue uses PQ scheduling (rarely on live networks) and all the other seven queues use weighted fair queuing (WFQ), the BE queue has the highest priority. If all the eight queues use WFQ scheduling, the priority is irrelevant to WFQ scheduling. For details about queue scheduling, see Queue Scheduling. Color Color, also referred to as drop precedence of packets on a device, determines the order in which packets in a queue are dropped when traffic congestion occurs. As defined by the Institute of Electrical and Electronics Engineers (IEEE), the color of a packet can be green, yellow, or red. Drop precedence is determined by the configured parameters. For example, if a maximum of 50% of the buffer size is configured for packets colored Green, whereas a maximum of 100% of the buffer size is configured for packets colored Red, packets colored Green have a higher drop precedence than packets colored Red. Packet priorities depend on the QoS configuration. Trusting the priority of received packets As described in section 2.2 Traffic Classification and Marking, after packets are classified on the DiffServ domain edge, internal nodes provide differentiated services for the packets that are classified. A downstream node can resume the classification result calculated on an upstream node or perform another traffic classification based on its own criteria. If the downstream node resumes the classification result calculated on an upstream node, the downstream node trusts the QoS priority (DSCP, IP precedence, 802.1p, or EXP) of packets received on the interface connecting to the upstream node. This is called the mode of trusting the interface. The switch trusts the following priorities: 802.1p priority The switch classifies packets based on 802.1p priorities and searches the mapping table of 802.1p priorities and CoS values. The switch classifies untagged packets based on the default 802.1p priority of an interface. Then the switch maps CoS values to 802.1p priorities and provides differentiated services. DSCP priority The switch classifies packets based on DSCP priorities and searches the mapping table of DSCP priorities and CoS values. Then the switch maps CoS values to DSCP priorities and provides differentiated services. The switch implements priority mapping according to the priority mapping table. In the DiffServ model, different DS domains allow different PHB mappings, so the device needs to allow an administrator to define DS domains and set different mappings in different DS domains. 18

24 Huawei switch allows an administrator to define DS domains. In addition, the system defines the default DS domain default. You can modify mappings in the DS domain default, but cannot delete it. Huawei switch provides the following priority mapping modes: DiffServ domain: S9700, S7700, S5700HI, S5710EI, S5710HI, and S6700 Priority mapping table: S5700SI, S5700EI, S5700LI, S5700S-LI, and S2750 If the mapping table is used, priority mapping implements mapping between packet priorities and PHBs, but cannot implement mapping between packet priorities and colors. All packets are green by default. When the DiffServ domain is used, run the display diffserv domain name default command to view the default mappings. When the mapping table is used, run the display qos map-table command to view the default mappings. The following tables show the mappings in the DiffServ domain: Table 2-5 describes the mappings from 802.1p priorities to PHBs and colors. Table 2-6 describes the mappings from DSCP priorities to PHBs and colors. Table 2-7 describes the mappings from precedences to PHBs and colors. Table 2-8 describes the mappings from EXP priorities in MPLS packets to PHBs and colors. Table 2-5 Mappings from 802.1p priorities to PHBs and colors 802.1p Priority PHB Color 0 BE Green 1 AF1 Green 2 AF2 Green 3 AF3 Green 4 AF4 Green 5 EF Green 6 CS6 Green 7 CS7 Green Table 2-6 Mappings from DSCP priorities to PHBs and colors DSCP PHB Color DSCP PHB Color 0-7 BE Green 28 AF3 Yellow 8 AF1 29 BE Green 9 BE 30 AF3 Red 10 AF1 31 BE Green 19

25 DSCP PHB Color DSCP PHB Color 11 BE 32 AF4 12 AF1 Yellow 33 BE 13 BE Green 34 AF4 14 AF1 Red 35 BE 15 BE Green 36 AF4 Yellow 16 AF2 37 BE Green 17 BE 38 AF4 Red 18 AF2 39 BE Green 19 BE 40 EF 20 AF2 Yellow BE 21 BE Green 46 EF 22 AF2 Red 47 BE 23 BE Green 48 CS6 24 AF BE 25 BE 56 CS7 26 AF BE 27 BE Table 2-7 Mappings from precedences to PHBs and colors IP Precedence PHB Color 0 BE Green 1 AF1 Green 2 AF2 Green 3 AF3 Green 4 AF4 Green 5 EF Green 6 CS6 Green 7 CS7 Green 20

26 Table 2-8 Mappings from EXP priorities in MPLS packets to PHBs and colors Exp PHB Color 0 BE Green 1 AF1 Green 2 AF2 Green 3 AF3 Green 4 AF4 Green 5 EF Green 6 CS6 Green 7 CS7 Green Complex Traffic Classification Traffic Classifier As networks rapidly develop, services on the Internet become increasingly diversified. Various services share limited network resources, so simple traffic classification can hardly meet requirements. Network devices must possess a high degree of awareness for services and support in-depth packet analysis to parse any packet field at any layer. Complex traffic classification meets the requirement to a certain degree. Complex traffic classification classifies packets in fine-grained manner based on rules such as the source MAC address, destination MAC address, inner and outer tags, source IP address, source port number, destination IP address, and destination port number. Complex traffic classification is deployed on edge nodes. Huawei switches provide various traffic classifiers and traffic behaviors. Traffic classifiers can be associated with traffic behaviors to form a traffic policy. To implement complex traffic classification, apply the traffic policy to an interface, a VLAN, or the system. The traffic policy based on complex traffic classification is also called class-based QoS. A traffic policy based on complex traffic classification is configured using a profile, which allows batch configuration or modification. A QoS profile defines the following items: Traffic classifier: defines a service type. The if-match clauses are used to set traffic classification rules. Traffic behavior: defines actions for classified traffic. Traffic policy: associates traffic classifiers with traffic behaviors. After a traffic policy is configured, apply it to an interface, a VLAN, or the system. A traffic classifier identifies packets of a certain type by using matching rules so that differentiated services can be provided for these packets. A traffic classifier can contain matching rules that do not conflict. If a traffic classifier has multiple matching rules, the AND/OR logic relationships between rules are described as follows: 21

27 Traffic Behavior OR: Packets that match any of the if-match clauses configured in a traffic classifier match this traffic classifier. AND: If a traffic classifier contains ACL rules, packets match the traffic classifier only when the packets match one ACL rule and all the non-acl rules. If a traffic classifier does not contain ACL rules, packets match the traffic classifier only when the packets match all the non-acl rules. On the Huawei switch, the default logic is OR and a traffic classifier can define matching rules based on the following items: Outer VLAN ID Inner and outer VLAN IDs in QinQ packets 802.1p priority in VLAN packets Inner 8021p priority of QinQ packets Outer VLAN ID or inner and outer VLAN IDs of QinQ packets Double tags of QinQ packets Destination MAC address Source MAC address Protocol type field encapsulated in the Ethernet frame header All packets DSCP priority in IP packets IP precedence in IP packets Layer 3 protocol type Inbound interface Outbound interface ACL rule Matching order of ACL rules After a traffic classifier is configured, the system matches packets against an ACL as follows: Checks whether the ACL exists (traffic classifiers can reference non-existent ACLs). Matches packets against rules in the order in which the rules are displayed. When packets match one rule, the system stops the match operation. An ACL can contain multiple rules and each rule specifies different packet ranges. ACL rules are matched according to the following matching modes: Config: ACL rules are matched according to the sequence in which they were configured. Auto: ACL rules are matched based on the depth-first principle. A traffic behavior is the action to be taken for packets matching a traffic classifier and is the prerequisite to configuring a traffic policy. Table 2-9 describes traffic behaviors that can be implemented individually or jointly for classified packets on a Huawei switch. 22

28 Table 2-9 Traffic behaviors Traffic Behavior Marking Traffic policing Traffic statistics Packet filtering Redirection Flow mirroring Description Sets or modifies the packet priority, such as 802.1p priority in VLAN packets and DSCP/internal priority in IP packets, to relay QoS information to the next device. Modifying packet priorities is also called re-marking. Limits network traffic and controls the usage of network resources by monitoring the traffic rate on a network. According to the configured traffic policing action, the device performs traffic policing for packets matching traffic classification rules, and discards excess packets or re-marks colors or CoS values of the excess packets. According to the configured traffic statistics action, the device collects statistics on packets matching traffic classification rules. The statistics on forwarded and discarded packets after a traffic policy is applied help you check whether the traffic policy is correctly applied and locate faults. Is the basic traffic control method. The device determines whether to drop or forward packets based on traffic classification results. Determines the packet forwarding path based on traffic classification results. According to the configured redirection action, the device redirects the packets matching traffic classification rules to the CPU, specified next hop address, or specified interface. The traffic policy that contains the redirection action can only be applied to the inbound direction of the system, an interface, or a VLAN. Copies the packets of an observed flow and then sends the copy to a specified observing interface. Usage Voice services, video services, and data services have QoS requirements in descending order of priority. On an enterprise network, an aggregation switch often connects to multiple access switches. You can configure traffic policing on the inbound interfaces of the aggregation switch to limit traffic. The enterprise NMS often provides this function, and monitors traffic based on services or users. It has the following functions: Limits resources accessed by some users. Filters out packets matching blacklist entries to protect the enterprise network. If there is a backup link in the outbound direction, configure redirection to the next hop address so that the device redirects high-priority services such as voice and video services to a higher-bandwidth or more stable link. Using this action, you can collect incoming and outgoing packets on an interface for fault analysis. 23

29 Traffic Policy You can apply a traffic policy bound to traffic behaviors and traffic classifiers to the system, an interface, or a VLAN so that the device can provide differentiated services. When creating a traffic policy on a Huawei switch, you can specify the matching order of traffic classifiers in the traffic policy. The matching order includes the auto order and config order. Auto order: The matching order depends on priorities of traffic classifiers. The traffic classifiers based on the following information are in descending order of priority: Layer 2 and Layer 3 information, Layer 2 information, and Layer 3 information. A traffic classifier with the highest priority is matched first. If the config order is used, traffic classifiers are matched in the sequence in which traffic classifiers were bound to the traffic policy. A traffic classifier that was bound to the traffic policy first is matched first. Applying a traffic policy globally Only one traffic policy can be applied to the system or slot (stack on chassis or box switches) in one direction. A traffic policy cannot be applied to the same direction in the system and slot simultaneously. In a stack composed of box switches, a traffic policy that is applied to the system takes effect on all the interfaces and VLANs of all the member switches in the stack. The system then performs traffic policing for all the incoming or outgoing packets that match traffic classification rules on all the member switches. A traffic policy that is applied to a specified LPU takes effect on all the interfaces and VLANs of the member switch with the specified stack ID. The system then performs traffic policing for all the incoming or outgoing packets that match traffic classification rules on this member switch. On a box switch in a non-stack scenario, a traffic policy that is applied to the system takes effect on all the interfaces and VLANs of the box switch. The system then performs traffic policing for all the incoming or outgoing packets that match traffic classification rules on the box switch. Traffic policies applied to the LPU and system have the same functions. When a traffic policy is applied globally on a chassis switch, the chassis switch performs traffic policing for all the incoming or outgoing packets that match traffic classification rules. When a traffic policy is applied to an LPU on a chassis switch, the chassis switch performs traffic policing for all the incoming or outgoing packets that match traffic classification rules. Applying a traffic policy to an interface On a Huawei switch, a traffic policy can be applied to only one direction on an interface, but a traffic policy can be applied to different directions on different interfaces. After a traffic policy is applied to an interface, the system performs traffic policing for all the incoming or outgoing packets that match traffic classification rules on the interface. Applying a traffic policy to a VLAN On a Huawei switch, a traffic policy can be applied to only one direction in a VLAN. Figure 2-9 shows relationships between an interface, traffic policy, traffic behavior, traffic classifier, and ACL. 24

30 Figure 2-9 Relationships between an interface, traffic policy, traffic behavior, traffic classifier, and ACL Interface Interface Policy Policy C&B C&B C&B Classifier Classifier Behavior Match Match Match Action Action ACL ACL Rule Rule Rule Rule The relationships are as follows: A traffic policy can be applied to different interfaces. One or more pairs of traffic classifiers and traffic behaviors can be configured in a traffic policy. A pair of a traffic classifier and a traffic behavior can be configured in different traffic policies. One or more if-match clauses can be configured in a traffic classifier, and each if-match clause can specify an ACL. An ACL can be defined in different traffic classifiers and contains one or more rules. One or more actions can be configured in a traffic behavior. Example: Configure two pairs of traffic classifiers and traffic behaviors in a traffic policy. acl 3001 rule permit ip source rule permit ip source acl 4001 rule permit vlan-id 10 rule permit source-mac traffic classifier 11 if-match acl 3001 traffic classifier 12 if-match acl 4001 Create a traffic policy and apply the traffic policy to an interface. traffic policy 1 classifier 11 behavior 11 classifier 12 behavior 12 (The traffic behavior configuration is not mentioned here.) 25